|
Posted by David Tiktin on January 25, 2008, 6:26 pm
If you were Registered and logged in, you could reply and use other advanced thread options
Did anyone happen to see the recent article by blogger Joel Spolsky
which in passing seems to advise disabling auto-negotiation in
"mission critical" environments?
http://joelonsoftware.com/items/2008/01/22.html
Paragraphs 7-9 contain the part I'm curious about.
I infer from the article that Joel's company disables auto-
negotiation as a matter of policy on all Ethernet equipment in their
data center. Is that a common practice? Would you recommend it?
(I know that there were compatibility problems with different
implementations when auto-negotiation was first introduced, but I
assume we're talking about "modern" equipment ;-)
BTW, the conclusion of the "Ethernet Autonegotiation Best Practices"
document from SUN Microsystems is a recommendation to leave auto-
negotiation enabled, but I'm not sure they are addressing the same
"mission critical" installation issue Joel Spolsky is talking about.
(See http://www.sun.com/blueprints/0704/817-7526.pdf)
I also wonder about the analysis that auto-negotiation was the cause
of the failure. There aren't any details in the article, but the
link in question was up for some time and then failed. If it was
working properly for a time, presumably auto-negotiation also worked
when the link came up. How could auto-negotiation have caused it to
fail later? (I'm not saying it couldn't. I just wonder how it could
have.)
I'd appreciate your comments.
Dave
--
D.a.v.i.d T.i.k.t.i.n
t.i.k.t.i.n [at] a.d.v.a.n.c.e.d.r.e.l.a.y [dot] c.o.m
|
 
| |
Posted by glen herrmannsfeldt on January 25, 2008, 6:50 pm
If you were Registered and logged in, you could reply and use other advanced thread options
David Tiktin wrote:
> Did anyone happen to see the recent article by blogger Joel Spolsky
> which in passing seems to advise disabling auto-negotiation in
> "mission critical" environments?
In the case where both ends are under the same administrator, and
locked in a wiring closet it might make sense. For ports that go
to user machines where ordinary users can move plugs around, I
would say no.
-- glen
|
|
Posted by Rick Jones on January 25, 2008, 7:19 pm
If you were Registered and logged in, you could reply and use other advanced thread options > David Tiktin wrote:
> > Did anyone happen to see the recent article by blogger Joel Spolsky
> > which in passing seems to advise disabling auto-negotiation in
> > "mission critical" environments?
> In the case where both ends are under the same administrator, and
> locked in a wiring closet it might make sense. For ports that go
> to user machines where ordinary users can move plugs around, I
> would say no.
I'm not sure I'd even do it in the first case Glen mentioned, but my
feelings there aren't strong.
If the article were written oh 5-10 years ago I might have grudgingly
agreed with it because much kit didn't do autoneg right, but "today"
unless someone is buying some really lame kit, autoneg should "just
work."
It is also the default on everything (?) which means if you did
hardcode, and forgot to configure it that way on a new peice of kit,
or someone resets that kit to factory defaults, you are in a realm of
Sorrow and Woe.
A bit long in the tooth now, since it talks about 100BT but it still
applies:
How 100Base-T Autoneg is supposed to work:
When both sides of the link are set to autoneg, they will "negotiate"
the duplex setting and select full-duplex if both sides can do
full-duplex.
If one side is hardcoded and not using autoneg, the autoneg process
will "fail" and the side trying to autoneg is required by spec to use
half-duplex mode.
If one side is using half-duplex, and the other is using full-duplex,
sorrow and woe is the usual result.
So, the following table shows what will happen given various settings
on each side:
Auto Half Full
Auto Happiness Lucky Sorrow
Half Lucky Happiness Sorrow
Full Sorrow Sorrow Happiness
Happiness means that there is a good shot of everything going well.
Lucky means that things will likely go well, but not because you did
anything correctly :) Sorrow means that there _will_ be a duplex
mis-match.
When there is a duplex mismatch, on the side running half-duplex you
will see various errors and probably a number of _LATE_ collisions
("normal" collisions don't count here). On the side running
full-duplex you will see things like FCS errors. Note that those
errors are not necessarily conclusive, they are simply indicators.
Further, it is important to keep in mind that a "clean" ping (or the
like - eg "linkloop" or default netperf TCP_RR) test result is
inconclusive here - a duplex mismatch causes lost traffic _only_ when
both sides of the link try to speak at the same time. A typical ping
test, being synchronous, one at a time request/response, never tries
to have both sides talking at the same time.
Finally, when/if you migrate to 1000Base-T, everything has to be set
to auto-neg anyway.
rick jones
--
web2.0 n, the dot.com reunion tour...
these opinions are mine, all mine; HP might not want them anyway... :)
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...
|
|
Posted by glen herrmannsfeldt on January 26, 2008, 3:07 am
If you were Registered and logged in, you could reply and use other advanced thread options
(snip, I wrote)
>>In the case where both ends are under the same administrator, and
>>locked in a wiring closet it might make sense. For ports that go
>>to user machines where ordinary users can move plugs around, I
>>would say no.
> I'm not sure I'd even do it in the first case Glen mentioned, but my
> feelings there aren't strong.
I probably wouldn't either, but I could see it as a possibility.
It would also have to be only me doing the administrating so that
I wouldn't consider anyone else messing things up. Mission critical
means lots of documentation, though. Good locks to keep others away
from things they shouldn't touch...
-- glen
|
|
|
Yahoo! 
Google 
Windows Live 
del.icio.us 
digg 
Netscape 
XML site map